Research On MEMS-Based Pedestrian Position Technology

In this paper,pedestrian location technology based on information fusion algorithms of low cost sensor is researched.According to the characteristics of pedestrian motion,the correction algorithm based on gait recognition is proposed.The main work of this paper is as follows:First,this paper introduces the background of the pedestrians positioning technology.This chapter presents the developmental status of MEMS inertial devices and the main technical routes currently applied to pedestrian positioning shortly.Aiming at the shortcomings of the existing positioning technology,the pedestrian location technology of the shoe-type low-cost MEMS sensor is proposed,and the three aspects of sensor error identification and calibration,gait recognition and filtering algorithm are studied.Second,the sensor output is analyzed and modeled in this chapter.First of all,this part introduces the definition of the coordinate system and the attitude angle which are commonly used in pedestrian positioning.Study of gait pattern of walking is also given so that characteristics of the sensor output are obtained by the walking experiment.According to the different phases of gait,the sensor output is modeled and the zero speed judging condition is confirmed.Third,the MEMS sensor error is analyzed and modeled.Aiming at the error type of the sensor,different schemes are adopted to calibrate and correct.The multi-position,ellipsoidal fitting method is used to correct the deterministic error and Allan variance analysis is utilized to analyze the random errors in inertial devices.Fourth,estimation of walking state.According to the characteristics of nonlinearity and continuity of motion state,the extended Calman filter algorithm is selected.In the stage of zero speed gait,using observed velocity and angular velocity information and heading information obtained by magnetometer to correct the system errors.Finally,several experiments are utilized to verify the algorithm.The accuracy of static and dynamic attitude calculation of the proposed algorithm is validated by using the dualaxis position turntable.The outdoor walking experiment is subsequently carried out to verify the positioning performance of the proposed algorithm.